Review papers with raw data transparency

Single-figure visual summary

Key claims (evidence links inline):

• Claim: Fz proteins are ωRNA-guided dsDNA endonucleases (biochemistry + TAM screens + in vitro cleavage) Biochemistry and TAM screens
• Claim: Fz can be reprogrammed for human genome editing (cell assays + optimization) Human-cell editing results
• Claim: Structural conservation with TnpB/Cas12 (cryo-EM 2.7 Å SpuFz1 ternary complex) 2.7 Å cryo-EM SpuFz1

Main strengths

• Multimodal evidence: phylogenomics, biochemical TAM screens, native small-RNA sequencing of RNPs, high-resolution cryo-EM, and human-cell editing—convergent methods increase confidence Multimodal dataset
• Open data: NGS (PRJNA982412), phylogenetic tree (iTOL), code for TAM screen on GitHub/Zenodo, and structural deposits—facilitates reproducibility.

Key limitations & blindspots

• Biological role in native hosts remains speculative: authors propose transposon-associated propagation (analogy to TnpB) but do not demonstrate in vivo function or fitness effects in native eukaryotes Function in native hosts unresolved
• Editing efficiency in human cells is modest (single-digit to low-double-digit % indels) and variable across orthologues and loci; off-target and cellular toxicity assessments are limited (no genome-wide off-targets reported), which is essential for translational claims.
• Most structural/functional mechanistic detail is from one orthologue (SpuFz1); generality across Fz1 vs Fz2 and diverse eukaryotic clades requires more structures/biochemistry.
• Potential conflict-of-interest/patent links: several authors are co-inventors and company founders—note for assessing commercial/translation incentives (declared in paper).

Detailed evidence map (claims → data)

1. Phylogenomics showing Fz distribution and origins. The authors mined AlphaFold models and NCBI sequences to produce 3,003 curated representatives and a phylogeny showing Fz1 and Fz2 clades emerging from distinct TnpB branches, consistent with at least two independent horizontal transfer events into eukaryotes Phylogenomic survey
2. ωRNA identification and RNP binding. Small-RNA sequencing (native S. punctatus and yeast-purified SpuFz1 RNPs) identifies 88–90 nt ncRNAs with a conserved ~75-nt scaffold and variable 14–21 nt guide tails; RNP pulldowns confirm ωRNA association with Fz proteins ωRNA discovery
3. Biochemical TAM screens & cleavage patterns. TAM-enrichment screens with randomized 8N libraries identify distinct TAM motifs per orthologue; in vitro cleavage mapping yields variable cut patterns (sticky/blunt) depending on orthologue TAM and cleavage mapping
4. Substrate specificity & collateral activity. SpuFz1 is TAM- and guide-dependent for dsDNA cleavage and shows no detectable collateral activity on ssDNA/dsRNA/ssRNA in their assays (important mechanistic distinction from some Cas12s) Substrate specificity
5. Human-cell editing & optimization. Wild-type SpuFz1, NlovFz2 and MmeFz2 produce locus-dependent indels (up to 11.8% reported for NlovFz2); activity is improved by ωRNA engineering (5' extensions, MS2 loop, trimming of flexible stem) and by substitution mutations (C310R/D487K/T513K = SpuFz1-v2) reaching ~18.4% indels at some loci Human editing optimization
6. Cryo-EM structure & mechanistic implications. 2.7 Å map/model shows ωRNA scaffold-stabilizing contacts, TAM interactions (REC/WED), and RuvC active-site architecture with catalytic residues coordinating two Mg2+ ions and possible catalytic water—supports RNA-assisted catalysis and evolutionary continuity with TnpB/Cas12 Cryo-EM mechanistic map

Critical interpretation & what would falsify the core claims

The central conclusion—that Fz proteins are programmable RNA-guided dsDNA endonucleases—is well supported by the congruence of: ωRNA identification from native loci, TAM dependence and specific cleavage in vitro, the high-resolution ternary structure showing an active RuvC site interacting with ωRNA and target DNA, and reporter genome-editing activity in HEK293FT cells after engineering.

What would disprove this paper's central claim?

• Failure to reproduce TAM-dependent cleavage in independent labs using purified RNP under the reported conditions;
• Genome-wide off-target assays demonstrating that observed indels are due to non-specific DNA damage from overexpression or cellular toxicity rather than guide-directed cleavage;
• Demonstration that ωRNA does not co-purify with Fz in native contexts or that small RNAs mapped are degradation products rather than functional guides.

Practical implications & recommended next experiments

• To assess translational potential: perform unbiased off-target mapping (GUIDE-seq / Digenome-seq / DISCOVER-seq) and cytotoxicity assays across cell types and delivery modalities (RNP, mRNA, AAV), including dose–response curves.
• To clarify biological roles: endogenous genetic perturbation (knockout/knockdown) of Fz loci and transposon copy-number assays in native eukaryote hosts (S. punctatus and others) to test the transposition-propagation hypothesis.
• To generalize mechanism: determine cryo-EM structures of representative Fz2 orthologues and TnpB comparisons to map conserved vs lineage-specific features; perform biochemical TAM mapping across more orthologues.

Key data & resources (from paper)

• DOI / paper: 10.1038/s41586-023-06356-2 Primary Nature paper
• Small RNA-seq / NGS: SRA BioProject PRJNA982412 (see Data availability) NGS deposit info
• Structure: PDB 8GKH; EMDB-40184 (cryo-EM map & coordinates) PDB/EMDB deposit
• Analysis code for PAM/TAM screens: GitHub and Zenodo links provided in paper for enrichment-analysis scripts.

Concluding critique — balanced view

The authors convincingly show that Fz proteins are ωRNA-guided DNA endonucleases and provide a structural rationale for guide recognition and TAM dependence. The combination of native ωRNA identification, TAM-biochemistry, and a high-resolution ternary structure is compelling evidence supporting the evolutionary and mechanistic claims. However, two major open areas remain: (1) the natural biological roles of Fz in eukaryotes (direct functional tests are missing), and (2) thorough specificity/toxicity profiling for genome-editing applications. For tool-development, additional engineering and delivery optimization with careful off-target profiling will be needed before translational claims can be assessed.